Tumor necrosis factor α (TNFα) is a cytokine produced by numerous cell types, including monocytes and macrophages, that was originally identified based on its capacity to induce the necrosis of certain mouse tumors (see e.g., Old, L. (1985) Science 230:630-632). Subsequently, a factor termed cachectin, associated with cachexia, was shown to be the same molecule as TNFα. TNFα has been implicated in mediating shock (see e.g., Beutler, B. and Cerami, A. (1988) Annu. Rev. Biochem. 57:505-518; Beutler, B. and Cerami, A. (1989) Annu. Rev. Immunol. 7:625-655). Furthermore, TNFα has been implicated in the pathophysiology of a variety of other human diseases and disorders, including sepsis, infections, autoimmune diseases, transplant rejection and graft-versus-host disease (see e.g., Vasilli, P. (1992) Annu. Rev. Immunol. 10:411-452; Tracey, K. J. and Cerami, A. (1994) Annu. Rev. Med. 45:491-503).
Because of the harmful role of human TNFα (hTNFα) in a variety of human disorders, therapeutic strategies have been designed to inhibit or counteract hTNFα activity. In particular, antibodies that bind to, and neutralize, hTNFα have been sought as a means to inhibit hTNFα activity. Some of the earliest of such antibodies were mouse monoclonal antibodies (mAbs), secreted by hybridomas prepared from lymphocytes of mice immunized with hTNFα (see e.g., Hahn T; et al., (1985) Proc Natl Acad Sci USA 82: 3814-3818; Liang, C-M., et al. (1986) Biochem. Biophys. Res. Commun. 137:847-854; Hirai, M., et al. (1987) J. Immunol. Methods 96:57-62; Fendly, B. M., et al. (1987) Hybridoma 6:359-370; Möller, A., et al. (1990) Cytokine 2:162-169; U.S. Pat. No. 5,231,024 to Moeller et al.; European Patent Publication No. 186 833 B1 by Wallach, D.; European Patent Application Publication No. 218 868 A1 by Old et al.; European Patent Publication No. 260 610 B1 by Moeller, A., et al.). While these mouse anti-hTNFα antibodies often displayed high affinity for hTNFα (e.g., Kd≦10−9 M) and were able to neutralize hTNFα activity, their use in vivo may be limited by problems associated with administration of mouse antibodies to humans, such as short serum half life, an inability to trigger certain human effector functions and elicitation of an unwanted immune response against the mouse antibody in a human (the “human anti-mouse antibody” (HAMA) reaction).
In an attempt to overcome the problems associated with use of fully-murine antibodies in humans, murine anti-hTNFα antibodies have been genetically engineered to be more “human-like.” For example, chimeric antibodies, in which the variable regions of the antibody chains are murine-derived and the constant regions of the antibody chains are human-derived, have been prepared (Knight, D. M, et al. (1993) Mol. Immunol. 30:1443-1453; PCT Publication No. WO 92/16553 by Daddona, P. E., et al.). Additionally, humanized antibodies, in which the hypervariable domains of the antibody variable regions are murine-derived but the remainder of the variable regions and the antibody constant regions are human-derived, have also been prepared (PCT Publication No. WO 92/11383 by Adair, J. R., et al.). However, because these chimeric and humanized antibodies still retain some murine sequences, they still may elicit an unwanted immune reaction, the human anti-chimeric antibody (HACA) reaction, especially when administered for prolonged periods, e.g., for chronic indications, such as rheumatoid arthritis (see e.g., Elliott, M. J., et al. (1994) Lancet 344:1125-1127; Elliot, M. J., et al. (1994) Lancet 344:1105-1110).
A preferred hTNFα inhibitory agent to murine mAbs or derivatives thereof (e.g., chimeric or humanized antibodies) would be an entirely human anti-hTNFα antibody, since such an agent should not elicit the HAMA reaction, even if used for prolonged periods. Human monoclonal autoantibodies against hTNFα have been prepared using human hybridoma techniques (Boyle, P., et al. (1993) Cell. Immunol. 152:556-568; Boyle, P., et al. (1993) Cell. Immunol. 152:569-581; European Patent Application Publication No. 614 984 A2 by Boyle, et al.). However, these hybridoma-derived monoclonal autoantibodies were reported to have an affinity for hTNFα that was too low to calculate by conventional methods, were unable to bind soluble hTNFα and were unable to neutralize hTNFα-induced cytotoxicity (see Boyle, et al.; supra). Moreover, the success of the human hybridoma technique depends upon the natural presence in human peripheral blood of lymphocytes producing autoantibodies specific for hTNFα. Certain studies have detected serum autoantibodies against hTNFα in human subjects (Fomsgaard, A., et al. (1989) Scand. J. Immunol. 30:219-223; Bendtzen, K., et al. (1990) Prog. Leukocyte Biol. 10B:447-452), whereas others have not (Leusch, H-G., et al. (1991) J. Immunol. Methods 139:145-147).
Alternative to naturally-occurring human anti-hTNFα antibodies would be a recombinant hTNFα antibody. Recombinant human antibodies that bind hTNFα with relatively low affinity (i.e., Kd˜10−7M) and a fast off rate (i.e., Koff˜10−2 sec−1) have been described (Griffiths, A. D., et al. (1993) EMBO J. 12:725-734). However, because of their relatively fast dissociation kinetics, these antibodies may not be suitable for therapeutic use. Additionally, a recombinant human anti-hTNFα has been described that does not neutralize hTNFα activity, but rather enhances binding of hTNFα to the surface of cells and enhances internalization of hTNFα (Lidbury, A., et al. (1994) Biotechnol. Ther. 5:27-45; PCT Publication No. WO 92/03145 by Aston, R. et al.)
Accordingly, human antibodies, such as recombinant human antibodies, that bind soluble hTNFα with high affinity and slow dissociation kinetics and that have the capacity to neutralize hTNFα activity, including hTNFα-induced cytotoxicity (in vitro and in vivo) and hTNFα-induced cell activation, are still needed.